We apply successfully a ChIP - Sybr-Green method in our group for analysis of TF binding enrichment within single promoters. Can one apply ChIP-Seq approach not for a genome-wide analysis but only for a 10 defined promoters simultaneously?
Maria- I don't understand why you would want to limit yourself to these 10 promoters. You don't need high-throughput sequencing for 10 promoters. Direct cloning and sequencing will work. If you want to know exactly where within the promoters your factor is bound, consider that even traditional ChIP-Seq has limited resolution (peaks might be 50-100bp wide). The best approach to find the exact motif detected by your factor is Exo-Seq.
I would suggest that it might be beneficial to go genome-wide in your study since you can learn a lot more with a relatively modest investment of resources.
not sure if this has been done, but I guess you can do thing with custom enrichment approaches as used, for instance, for exome sequencing... But it would be a tremendous overkill of data I guess... Maybe you'd be better of with tiling arrays spanning your loci and do ChIP on chip..
Daniel, I though exo-seq is confined to exons. How does this get at TF binding? Most TF binding sites are too short and degenerate to be able to enrich for. I would agree that if you are going to do high throughput, you might as well get as many reads as possible.
exo-seq is indeed limited to exons, but you can get custom made enrichment thingies for any sequence you want. Question is I think if it's worth it for just 10 promoters...
I don't see the point in using ChIP-seq if you are just interested in 10 genes. One could probably combine ChIP with sequence capture, but that would be prohibitively expensive. And the coverage on each loci would be insanely high...
Never tried it myself or konow of anybodx who did it but if you for a particular reason don't want to use qPCR, the NanoString technology could be an interesting alternative (see: http://www.nanostring.com/products/ChIP-String).
Well, the thing is - I just don´t feel myself confident enough to interprete and evaluate genome-wide ChIP-Seq data, so I thought to narrow first the number of promoters. But, surely, you are right - it is probably a bit unclever way to proceed.
With ChIP-chip I´m a bit sceptical how well each promoter is represented, how tight are the probes positioned.
In any case, we would be happy to collaborate with some experts in ChIp-seq in respect of analyzing binding enrichment of a particular TF in human primary material.
Hi guys, I meant ChIP-Exo, where exonucleases are used to trim off any DNA sequences that are not covered by the TF of interest. Sorry for any confusion this caused!
Maria- the analysis of ChIP-Seq datasets is remarkably easy to learn. The "hard" part is making sure you have a good ChIP-grade antibody and can recover a DNA library with sufficient complexity to get meaningful genomic coverage. In your qPCR, what is your enrichment at promoters of interest relative to input and relative to negative control regions? If you can see enrichment of close to 1% of input or higher, you're in good shape.
Daniel - I can imagine it depends on the background of a scientist. But of course one could try. Yes, you´re right, we were indeed searching for good antibodies. Found one and we do get an enrichment around 1%, so this part is already optimized.
Another point is a "cost factor": ideally we would like to investigate the binding profiles of one factor within 10-12 defined gene promoters under different physiological conditions. Now if we go for ChIP-Seq genome-wide analysis in 10 donors plus different conditions - that could be a bit expensive, right?
Diego - we use routinely ChIP-qPCR (detetcting with Sybr-Green) for single promoters. I thought of applying kind of "middle-high-throughput" technology - smulatenous generation of binding profiles within 10-12 particular promoters of one gene group.
Another important consideration for ChIP-Seq is how your factor is recruited to DNA. Does it bind DNA directly? Is it motif dependent? ChIP-Seq works best when the answer to both questions is yes. If no, then you have reason to proceed with caution (it can be done, but it is harder to validate the results).
But you have to think of i) the diversity of the "specific" regulatory elements (RE) sequences present in the promoters [even if you selected only 10-12 for some good reasons, I presume] and ii) the diversity of the transcription factor (TF) you are evaluating since it might be a member of a family of TF. The hierachy of interactions is based on the diversity of these two components.
@Catherine- it is true that factors can bind different motifs, and the similar motifs can bind different factors, but I'm not sure how this would affect Maria's study. She already has selected a factor to study, and by using ChIP, she has the ability to focus on a specific family member provided the primary antibody employed can discriminate between related TFs.
If you are suggesting that her factor might bind different motifs in her 10 promoters or in a ChIP-Seq dataset, I agree. However, this isn't necessarily a problem.
I presume that the selection of the different genes and promoters was based on the presence (or not as control) of these RE Maria is interested in. In addition, one important point also is when were performed these expreiments since the concentration of the TF varies. If the concentration is too low, and if the affinity between the RE and the TF is not very good she might lose some ChIP interactions that she will be able to see at a time when the concentration of TF is increased.
Dear colleagues, thank´s a lot for sharing your thoughts - it helps indeed to hear experts opinion.
As a matter of fact, the choice of gene promoters was based on both "inducibility" effect and presence of predicted REs. Whereas activation of TF gives quite a pronounced effect on upregulation of transcription of downstream genes, in silico analysis of promoter sequence (10.000 bp upstream) results in a number of potential binding sites with different affinities. The second part is not really conclusive, playing around with stringency I can get from 2 to 100 potential sites of different scores within one promoter. The response elements can have mismatches.
That´s why my way of thinking was to screen the binding profiles of our TF using ChIP-Seq in the "responsive" gene group - if we do not see any binding, we can think of a secondary effects, if we manage to define the peaks - we can move further for validation.
Maria- if I understand you correctly, you have already identified your TF of interest and feel confident that its binding is dependent on a particular, known, DNA motif. The problem is that the motif occurs multiple times in the candidate promoters, right?
In many cases, factors that strongly induce a target gene will bind multiple sites in the promoter. So in some cases, the prediction of several candidate sites (although probably not exceeding 5-10), is potentially relevant. If you want to find out which of these are function, there are a couple of strategies you could use.
1) genome-wide ChIP-Exo
2) if you are working in mouse or human cells, use the ENCODE datasets at UCSC to look at known DNase HSS sites, and known TF binding sites. TFs tend to bind in groups, so if your motif maps to a known DNAse HSS site or an ENCODE TF cluster, it is more likely to be functional than those that do not.
3) clone some promoters of interest into a luciferase reporter and systematically mutagenize your motifs of interest (although this is not possible for dozens of sites or all your promoters at once).
Perhaps #3 is best suited to your system, since it sounds like you are trying to determine whether a particular TF is responding to a stimulus given to your cells. In this case, you might even perform a deletion series in order to map which portion of the promoter responds to your stimulus of interest, and see if the motif is also present in this region, and subsequently whether it is required (effect of a targeted mutation).